Rapid identification model for yak milk and milk doped with yak milk

Abstract

The invention belongs to the technical field of dairy product analysis and particularly relates to a rapid identification model for yak milk and milk doped with the yak milk. The invention relates to the field of mid-infrared spectrum analysis. The method comprises the following steps of 1) taking fresh yak milk and a yak milk sample doped with milk; 2) scanning the sample in the medium infrared spectrum range to obtain medium infrared spectrum data; 3) preprocessing the original mid-infrared spectrum to remove abnormal values; 4) dividing the preprocessed data set into a training set and a test set according to a stratified sampling principle; 5) screening a spectral band for modeling; 6) combining different spectrum preprocessing methods and modeling algorithms, establishing an identification model, evaluating the model by using accuracy and a kappa coefficient, and screening out the combination of the preprocessing method and the modeling algorithm with the optimal effect to obtain an optimal model; and 7) verifying the model, and evaluating the generalization ability of the model.

Description

technical field [0001] The invention belongs to the technical field of milk product analysis, and in particular relates to an identification method for yak milk and yak milk mixed with milk. Relevant to the field of mid-infrared spectroscopy. Background technique [0002] Mahood et al. (2017) used Fourier-transform near-infrared spectroscopy combined with chemometrics to analyze the R values ​​of the PLS-DA model established for pure yak milk and yak milk mixed with 10% milk. 2 is 97%, and the root mean square error is 0.08%. The model performs second-order differential preprocessing on the spectrum to distinguish overlapping peaks and eliminate baseline changes. The cross-validation method is the leave-one-out method, which is suitable for the verification of small samples [5] . Hu Xiaoxiao (2020) added different concentrations of non-dairy creamer, dextrin, whey powder and soybean protein isolate to yak milk powder, and established the SIMCA classification model and PLS ...

AI Technical Summary

Problems solved by technology

The data output by the mid-infrared spectrometer is a matrix of n×1060 (n is the sample size), the data is huge, and it is difficult to avoid data incompleteness, inconsistency, and is easily disturbed by noise (errors or outliers). Low-quality data will lead to Poor data mining results, so some method is needed to preprocess the output data

Images